False-Positive Recall and False-Positive Biopsy Rates in Mammography Screening: A TOSYMA Trial Subanalysis

Background False-positive results may negatively affect participants and diminish the efficiency of mammography screening programs. Purpose To assess whether the higher breast cancer detection rate with digital breast tomosynthesis (DBT) plus synthesized mammography (SM) compared with digital mammography (DM) in the randomized TOmosynthesis plus SYnthesized MAmmography (TOSYMA) trial comes with higher false-positive rates. Materials and Methods The TOSYMA trial was conducted from July 2018 to December 2020 in women aged 50-70 years (mean age, 57.7 years ± 5.6 [SD]), with 1:1 allocation to the DBT plus SM or DM study arm (NCT03377036). True- and false-positive recall rates, false-positive biopsy rates, and positive predictive values of recalls and biopsies performed were compared per arm. Women with false-positive findings were followed up for 24 months. The analyses were strictly exploratory. Results There were 49 762 women in the DM arm and 49 715 in the DBT plus SM arm. The recall rate was 48.6 per 1000 women screened (2416 of 49 715) with DBT plus SM and 49.9 per 1000 women screened (2483 of 49 762) with DM. The true-positive recall rate of DBT plus SM (8.4 per 1000 women screened; 416 of 49 715) was higher than that of DM (6.2 per 1000 women screened; 306 of 49 762), a difference of 2.2 per 1000 women screened (CI: 0.4, 4.0). The false-positive recall rate was lower with DBT plus SM (40.2 per 1000 women screened; 2000 of 49 715) than with DM (43.7 per 1000 women screened [2177 of 49 762]; difference: -3.5 per 1000 women screened [CI: -7.8, 0.7]), particularly in the first screening round (DBT plus SM: 94.0 per 1000 women; DM: 109.6 per 1000 women screened; difference, -15.6 per 1000 women screened [CI: -31.0, -0.2]). The positive predictive value of recall was higher with DBT plus SM (17.2%; 416 of 2416 women recalled) than with DM (12.3% [306 of 2483 women recalled]; difference, 4.9% [CI: 1.5, 8.3]). The false-positive biopsy rate (7.8 per 1000 women screened; 388 of 49 715) was higher with DBT plus SM (1.8 per 1000 women screened; CI: 0.0, 3.6), and the positive predictive value of biopsies performed was 51.7% (416 of 804 women with biopsies performed; difference, 1.1%; CI: -7.9, 10.2). Interval cancers after false-positive recalls were rare (five with DBT plus SM and two with DM). Conclusion This exploratory analysis of the large randomized TOSYMA trial indicates that DBT plus SM showed lower false-positive recall rates than DM, especially in first-round screening participants. Mildly raised false-positive biopsies of DBT plus SM were compensated by more detected breast cancers. ClinicalTrial.gov identifier: NCT03377036 © RSNA, 2025 Supplemental material is available for this article. See also the editorial by Mesurolle in this issue.